According to Shing-Fat Fred Ma:
> Thanks for your explanations, Grant.  Though some it raises
> more questions, in case you or anyone else care to elaborate.
>
> Grant McDorman wrote:
>
> > -depth does not control the connection depth; instead, it tells the
viewer
> > to try to create its local window at that depth *with the default
visual*.
>
> I thought the whole point of specifying the veiwer depth was to
> reduce bandwidth.  Why would one want to reduce the viewer
> depth without also having it reflected in the communication?
> For the thin client model, the (naively) reasonable thing to do
> is to get server to quantize each color plane to the coarseness
> requested by the viewer depth.  Obviously something I'm
> missing here.

Viewer depth is to adapt the viewer to the local machine (that is, the
machine on which the viewer is displaying). It is *not* to control
bandwidth, although it may affect that as a side effect. Options which
control the viewer display are -depth, -truecolour, -owncmap, and
-fullscreen (and the standard Xt option -display).

Conversely, the bandwidth control options are -encodings and -bgr233 (and,
for TightVNC, -compresslevel and -quality).

The viewer will force -bgr233 if the remote system - the VNC server - is
TrueColour and the local system is 8-bit (pseudo-colour).

> > What is happening in your case is that your local system's default visual
is
> > 8-bit pseudo colour, and a 24-bit pseudo-colour visual doesn't exist; the
> > viewer, in that case, will use the only available pseudo-colour visual
> > (8-bit) and behave as if the -bgr233 command line option was supplied.
>
> This would certainly explain the "truant" viewer behaviour,
> though I was under the impression that the display is *not*
> pseudocolor unless the server is started with "-cc 3".  I ran
> into a problem with CAD tools for IC layout because it wanted
> a pseudocolor visual, and someone on the VNC list rightly
> suggested that the "-cc 3" was required.  I'm assuming that
> "visual" is a property of the server display only, in particular
> the display created by vncserver.

Here I am talking about the local display, not the VNC server. The VNC
viewer can display using TrueColour if the local display supports it;
however, if it is not the default you have to tell the viewer explictly to
use it.

One must remember that, in this case, there are *two* X servers involved:
the VNC (virtual) server and the physical X server (presumably XSun) that
you use to run the viewer. In order to run and view an application in
TrueColour, *both* must support TrueColour. If either is 8-bit, then you
will get an 8-bit palette of some form.

The vncviewer -truecolour and -depth 24 options can be used to view *any*
VNC server configuration; i.e. you can view an 8-bit PseudoColour, or an 8,
16, or 24 bit TrueColour VNC server.

To further clarify: Most X servers, including the VNC server, will only
support a single visual (and depth) at one time. Some, including the Sun X
server, will support multiple simultaneous visuals; one of these will be the
default visual. The VNC viewer '-depth' and '-truecolour' options are only
useful when it is displaying on a (client) X server like this that have
multiple visuals and/or depths.

> > If the server has a true colour visual available, this command line
should
> > do what you want:
> >   vncviewer -depth 24 -truecolour _server_:_num_
>
> Just something I want to clarify here.  By server, I presume
> you mean vncserver rather than host machine for the vncserver?
> Because for VNC, it seems that server and display and visual
> are all one and the same.  A server cannot have more than 1
> display, some being true color while others are not.  Assuming
> they are synonymous, then it seems the existence of "-cc 3"
> and "-truecolour" specify the same thing, except one is on the
> server side and the other is on the viewer side....right?

No, sorry. This, again, is the viewing machine (a Sun machine, if I recall
correctly). -truecolour has no effect on the VNC server whatsoever. "server"
here was "local X server" (or $DISPLAY where the VNC viewer is run).


--

Grant McDorman <[EMAIL PROTECTED]>, Sr. Software Design Consultant
Cedara Software Corp.  <URL:http://www.cedara.com>
  (formerly I.S.G. Technologies Inc.)
Mississauga, Ontario, Canada

[demime 0.99c.8 removed an attachment of type application/pgp-signature]
_______________________________________________
VNC-List mailing list
[EMAIL PROTECTED]
http://www.realvnc.com/mailman/listinfo/vnc-list

Reply via email to